[R] How to code the factor for ANOVA

Jinsong Zhao jszhao at yeah.net
Mon Jun 26 07:05:17 CEST 2017


Hi there,

I have a experimental design related question. I have done a experiment 
with 3 factors. The design matrix is similar to:

> data.frame(Factor.1 = rep(rep(0:2, each = 3),3), Factor.2 = 
rep(c("U","S","N"), 9), Factor.3 = rep(0:2, each = 9))
    Factor.1 Factor.2 Factor.3
1         0        U        0
2         0        S        0
3         0        N        0
4         1        U        0
5         1        S        0
6         1        N        0
7         2        U        0
8         2        S        0
9         2        N        0
10        0        U        1
11        0        S        1
12        0        N        1
13        1        U        1
14        1        S        1
15        1        N        1
16        2        U        1
17        2        S        1
18        2        N        1
19        0        U        2
20        0        S        2
21        0        N        2
22        1        U        2
23        1        S        2
24        1        N        2
25        2        U        2
26        2        S        2
27        2        N        2

The Factor.2 indicates the type of a fertilizer, such as urea, and the 
Factor.3 indicates the weight of the fertilizer. So the treatments with 
Factor.3 == 0 at each levels of Factor.1 are same. That means No.1, No.2 
and No.3 are same, No.4, No.5 and No.6 are same, and No.7, No.8 and No.9 
are same. Thus, in the real experiment, those treatments were not 
performed as stated in the design matrix. For instance, only No.1, No.5 
and No.9 are performed.

In the case, how to code the factor for ANOVA? I really appreciate any 
comments and suggestions. Thanks in advance.

Best regards,
Jinsong



More information about the R-help mailing list